The general objectives of this project are to (1) assimilate data from language pathology to psycholinguistic theories of normal language processing, using aphasia as a testing ground for those theories; (2) develop assessments and analyses that can be used reliably across patients and across laboratories to yield the type of data suitable for (1); (3) obtain data that can be applied in the rehabilitation of aphasic disorders; and (4) develop computational models that instantiate and test specific accounts of psycholinguistic disorders and their relation to the normal language system. In this project period, our efforts will be aimed at three areas: (a) syntactic disorders, where we will (i) complete a large scale study of sentence production in nonfluent aphasics that seeks to confirm previously reported dissociations of morphological and structural indices and (ii) conduct on-line studies of sentence processing to test competing accounts of the deficit that underlies asyntactic comprehension; (b) lexical-semantic deficits, where we will (i) explore in detail the disorder that selectively compromises knowledge of man-made artifacts, seeking to establish its implications for naming and object use, as well as its neuroanatomic substrate; and (ii) investigate the role of semantic complexity/and or specificity in the retrieval of verbs, using a story completion paradigm; (c) lexical- phonological retrieval, where we will (i) expand our effort to simulate aphasic naming patterns with an implemented interactive activation model, and to test predictions regarding other language functions that are derived from the model; (ii) use an error-elicitation procedure to investigate between-word sublexical interaction errors in fluent aphasic speech, including serial order effects, similarity constraints, and lexical bias; and (iii) further develop the model to deal with words of complex shape and multiple-word utterances.